China's Provisions on the Management of Multi-Channel Distribution Services for Internet Information Content took effect on 1 September 2026. The Cyberspace Administration of China (CAC), the Ministry of Public Security, the Ministry of Culture and Tourism, the State Administration for Market Regulation and the National Radio and Television Administration issued them on 29 May 2026 as Order No. 23. They add a new legal layer between platforms and creators: the multi-channel network (MCN), the agency that plans, produces, markets and brokers content for online creators. Under the rules, that agency carries duties of its own.
The strongest case for the rules
Regulators have a real problem to point at. The CAC's official Q&A says some institutions and platforms, "driven by traffic interests," hype trending events, fabricate personas and spread illegal content. It also names deepfakes, fake engagement metrics and counterfeit goods. Metric-faking and deceptive marketing harm audiences and honest advertisers in any country. Agencies that run hundreds of accounts also have more influence than any single creator, so it is arguable that they should carry duties proportionate to that influence. Some of the rules are ordinary consumer protection: MCNs must hold proper business registration, employ a dedicated content-management lead, and may not provide livestream publishing services to anyone under 16. Those are defensible.
What the rules require
The official text sets out a chain of duties, each one enforceable against a different party.
- Platforms must verify an MCN's qualifications before onboarding it and sign an onboarding agreement. Article 8 requires them to file the MCN's details with the provincial cyberspace department within 30 working days of onboarding.
- Platforms must also display, prominently, the name of the MCN behind each signed public account (Article 11).
- MCNs must verify the identity of the people who produce and run each account they manage (Article 13). They must also warn, suspend service to, or terminate creators who violate the law or their agreement (Article 20).
- Penalties under Article 26 include fines of 10,000 to 100,000 yuan. The CAC's summary describes a higher band of 100,000 to 200,000 yuan for serious harm.
This is intermediary liability arranged as a cascade. The state holds the platform responsible for the MCN, the MCN for the creator, and every link must be able to name the one below it. Compare this with the safe-harbour logic that underpins much of the open internet, where intermediaries are generally not liable for user content until they have notice of specific illegality. China's approach is closer to a licensing-and-attribution model. The state does not wait for notice, because it requires every account to have an identified, accountable operator in advance.
Why the design worries innovation and speech
Three features stand out, and none depends on the intentions of the drafters.
First, the rules make private agencies the enforcement point. When an MCN must suspend a creator over a violation of law or of platform rules, the agency's incentive is to over-remove. Its own exposure includes fines, loss of its filing, and reputational harm with the platform. A creator's income is not part of that calculation. Comparative research on notice-and-takedown regimes repeatedly finds this pattern, and the EFF's recent analysis of Brazil's new duty-of-care obligations warns of the same incentives toward over-censorship of protected speech. In China the pressure is stronger, because liability attaches to the whole chain and not to a single notified item.
Second, the standards are open-ended. The China Media Project, reviewing the January 2025 draft, noted that the rules require adherence to "correct political direction" and "public opinion guidance," language it links to older propaganda controls. That analysis concerned the draft, and I could not confirm from the final text how much of that language survived. If any of it did, agencies cannot know where the line is, and the rational response is caution.
Third, identity binding erodes the space for pseudonymous and independent voices. Verified identity for accounts that earn money is a reasonable fraud-prevention measure. The concern is that the same verified chain also serves as a map for enforcement, running from platform to agency to individual.
A proportionate alternative
The legitimate goals here are honest metrics, truthful advertising, child safety and transparency about who is behind an account. All of them can be met with narrower tools:
- Disclosure of commercial relationships between agencies and creators, which serves audiences without gating who may speak.
- Liability for specific, defined harms, such as fraud, counterfeit goods and exploiting minors, instead of broad duties over content.
- Judicial or independent review before an account can be terminated, with a right of appeal for creators.
- Clear, published definitions of prohibited conduct, so that agencies can comply without guessing.
Disclosure of the MCN behind an account is the kind of transparency measure that works in open markets, where an audience can weigh a brand tie for itself. In China's system, disclosure sits alongside filing with provincial regulators and mandatory suspension duties, so it operates as a control point and not as consumer information.
What to watch
The practical test is enforcement in the first months. Watch for three things: how many platforms publish their filing and annual-reporting procedures, whether provincial cyberspace departments publish enforcement notices, and whether agencies respond by dropping smaller creators whose compliance cost outweighs their revenue. Consolidation toward large, well-resourced MCNs would be the predictable result, and it would raise the barrier to entry for new voices. Other jurisdictions writing intermediary rules, including those adopting duty-of-care models, should note that layering liability through several intermediaries changes who speaks, and it does so before any individual case of harm arises.